News for hackers. Stuff that matters.
Re:How I didn't become a SoundClouder by hardwaresofton on Monday July 10, 2017 @07:17AM (#14734898)
While I don't think a 20 hour project is reasonable -- I do like the option of doing a shorter project when offered in lieu of a timed coding test or an hour interview.

I think it's one of the most close-to-actual-work things you can do (especially if they give you a problem at least close to what you're going to be working on everyday).

Designed well, I could imagine a test that would take 1 to 2 hours (of course you can't control how long someone will work on something, but if I ask you to reverse a string in python I don't think it's reasonable for that to take hours), and succinctly test for clean architecture and good general software engineering principles.

Re:How I didn't become a SoundClouder by zdragnar on Monday July 10, 2017 @01:04PM (#14737714)
In my experience designing these things, 1-2 hours only provides enough time for very simple assignments, which aren't enough to actually inform the reviewer much at all about you as an individual.

It's hard to get one at 4 hours, but that provides just enough to really see something while still being respectful of candidates' time.

Of course, this is entirely dependant on the type of programming that you're hiring for!

Re:How I didn't become a SoundClouder by pjmlp on Monday July 10, 2017 @07:33AM (#14734972)
The problem is that most of the time is like solving newspaper puzzles.

Last time I got a stupid exercise about counting amount of 4s and 0s in a number, while obeying to a specific pattern, then generate numbers that according to that pattern would match a specific formula, while taking into consideration some GCD relations among those numbers.

Yeah, cool exercise for a weekend. Zero value when writing enterprise distributed systems with web/native GUIs.